Hmmm. If we for example put all YouTube videos on the blockchain, it's clearly going to be enormous. I don't think it's practical to have 10000s of identical copies in that case.
...So we'd have to split it up in some way, and maybe store a handful of copies of every file distributed across the network.
If blockchain works exactly as now, yes. But what if it was only partially duplicated, with perhaps 1000 full copies in total rather than one for each node (perhaps millions)?